make 'includes' and 'does not include' work as substring tests

[cgrayson]

it looks like includes and does not include could never behave differently than is equal to and is not equal to. As in this test.

Discussion.

[alexj]

We'd like the ability to check whether a string contains another string, so we should ensure that includes and does not include work in this manner instead of removing the operators from the string type. Or am I missing something?

[cgrayson]

I think that's a separate issue. Currently, at least, includes and does not include have specific meaning with regard to arrays. Though they seem similar, includes is explicitly written to not do pattern matching in that context, so I'd be worried about "overloading" the operator like that.

For example, ['apple', 'beetle', 'carpenter'] includes 'bee' would be false (but that array includes 'beetle' would be true).

Maybe it would be fine. Maybe there should be a contains or matches instead. Maybe Alex W. has it all figured out & half written. 😃

[alexj]

That makes sense. contains would make more sense as a label too. Thanks Chris.

[alexkwolfe]

I also think what you've said makes sense, Chris. 👍

[cgrayson]

After further discussion… 😊 rather than remove these, Wolfe & I decided it would be better to make them operate as substring tests (i.e., as Jones mentioned in the first place), after all.

These still won't be full regex operators, though; that would still be a separate operator (say, matches).

For the record we also talked about making this string operator have a distinct name other than the array operator (includes), such as contains. There's potential user confusion, when working with a rule, if they're offered operators that don't match what they're expecting based on data type. In the end we think it will be less confusing to have includes mean similar things with strings & arrays, instead of having different, synonymous names.

[cgrayson]

Looking at this now, I screwed up which issue should be open or closed. The valid string operators are listed here, but the way they work is defined in leadconduit-rules. So, as I should have done on Sep 2, I'm: closing this, and reopening rules #26.